Regional community-acquired urinary tract infections in Israel: diagnosis, pathogens, and antibiotic guidelines adherence: A prospective study

International Journal of Infectious Diseases (2007) 11, 245 250 http://intl.elsevierhealth.com/journals/ijid Regional community-acquired urinary tract infections in Israel: diagnosis, pathogens, and antibiotic guidelines adherence: A prospective study L. Nesher *, V. Novack, K. Riesenberg, F. Schlaeffer Department of Internal Medicine, Soroka Medical Center and Ben-Gurion University, Box 151, Beer Sheva, 84101, Israel Received 19 November 2005; received in revised form 7 March 2006; accepted 9 March 2006 Corresponding Editor: Andy I.M. Hoepelman, Utrecht, The Netherlands KEYWORDS Urinary tract infections; Diagnosis; Anti-bacterial agents; Practice guidelines Summary Introduction: The identification and treatment of hospitalized patients with communityacquired urinary tract infections (CAUTI) may be a challenge. The pathogens causing the infection and their relative proportions vary geographically and with time. This observational prospective study had three primary goals: (1) to estimate the likelihood of diagnosis of CAUTI upon admission; (2) to evaluate adherence to the institutional recommendations; (3) to assess the compatibility of the current local antibiotic recommendations with a pathogen s distribution and with its drug sensitivities. Methods and results: Two hundred and twenty-three patients with positive urinary cultures fulfilling criteria for CAUTI were studied. Only 54 (24.2%) were diagnosed as having a urinary tract infection upon admission. Approximately 90% of the patients, who were correctly diagnosed, received the institutional recommended antibiotic therapy (ofloxacin or cefuroxime). Gramnegative intestinal flora comprised 86.1% (192 patients) of the causative microorganisms. Of these, 20.3% of the pathogens demonstrated resistance to ofloxacin and 19.8% to cefuroxime. The prevalence of Escherichia coli, the most common pathogen of UTI, significantly declined in the current study, from 70.5% in 1991 to 56% in 2000. Conclusions: We observed a low sensitivity in diagnosing community-acquired urinary tract infections upon admission. In patients correctly diagnosed, the use of recommended antibiotics was high. A substantial percentage of the pathogens were resistant to the recommended antibiotics. This study stresses the need for frequent re-evaluation of the prevalence of pathogens involved in regional community-acquired urinary tract infections and the adjustment of the empirical first-line treatment accordingly. # 2006 International Society for Infectious Diseases. Published by Elsevier Ltd. All rights reserved. * Corresponding author. Tel.: +972 8 6400235; fax: +972 8 6403366. E-mail address: nesherke@bgu.ac.il (L. Nesher). 1201-9712/$32.00 # 2006 International Society for Infectious Diseases. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijid.2006.03.004

246 L. Nesher et al. Introduction Urinary tract infections (UTI) are the most common form of bacterial infections, affecting people throughout their lifespan. 1,2 The term UTI covers a variety of clinical entities, ranging from cystitis and prostatitis to pyelonephritis. UTI are more common in females, except in the elderly. As many as 50 80% of women experience at least one episode of UTI during their lifetime, most of which present as uncomplicated cystitis. 1 3 In a recent prospective cohort study of sexually active healthy women, the incidence of acute cystitis was 50 70 episodes per 100 person-years. 4 Years of antibiotic over-prescription and abuse on the one hand, and a decline in the development of novel antibiotics on the other, have led to a tendency among physicians to shy away from prompt and aggressive prescription of these drugs. 5,6 This is especially the case with broad-spectrum antibiotics. The predicted changes in pathogens and in their occurrence makes it highly advisable that empirical, first-line antibiotic treatment should be reviewed periodically in every regional tertiary medical center. 5,7 9 To this end, rigorous, periodical re-evaluation of regional protocols for empirical treatment of community-acquired ascending urinary tract infections (CAUTI) should be performed in every major community or tertiary medical center, and its conclusions used as guidelines for that institution as well as for primary clinics and secondary community centers. 5,6,9,10 Longitudinal time studies may give important indications on changes in trends of the common pathogens and provide improved early tools to combat them. A CAUTI, especially when complicated, may be a potentially life-threatening condition; 2,10,11 therefore adequate early empiric treatment is essential and may prevent or reduce morbidity and mortality. 2,11 Early diagnosis of patients on the more severe side of the spectrum of CAUTI (pyelonephritis) is difficult and many times is done only after obtaining results from urine or blood cultures, and in some cases not made at all. 11,12 Hence, it is critical to reassess the initial, empirical treatment and its efficacy with reference to the patient status. 9,13,14 The pathogens causing UTI may be found along the entire urinary tract and therefore it is accepted that the entire tract be considered as a single afflicted entity. 2 Patients hospitalized with CAUTI are usually regarded as severe cases and are in need of intravenous antibiotic therapy. 2,3,9 Recent studies have demonstrated that the epidemiology of bacteria causing CAUTI varies from country to country and even between regional hospitals. 5 7 Therefore, our observational prospective study was designed: 1. To determine the frequency of CAUTI diagnosed as such, prior to a positive urinary culture. 2. To determine the proportion of CAUTI cases that received a first-line antibiotic according to current local recommendations. 3. To assess the incidence of the most common uropathogens in our local cases of CAUTI. 4. To assess the compatibility of present local antibiotic guidelines and the level of adherence, with the prevalence of bacteria in our region. Materials and methods The study was performed at the Soroka University Medical Center, a 1000-bed university-affiliated referral center in Southern Israel, serving a population of 750 000. The study was conducted prospectively throughout 2000 and included hospitalized patients meeting the following inclusion criteria: (1) oral body temperature of 38.2 8C; (2) at least Figure 1 Diagnostic steps in this study of hospital UTI (number of patients).

Regional community-acquired UTI in Israel 247 one of the following symptoms: dysuria, frequency, urgency, flank pain, costovertebral angle tenderness; (3) positive urine culture (over 10 5 CFUs) of a single pathogen obtained upon admission. 12,13 A complicated urinary tract infection was defined as CAUTI on admission in males, pregnant women, diabetics, immunosuppressed patients, or in patients with a structural or obstruction abnormality of the genito-urinary tract. 3,13 A patient was categorized into the dementia group if he was unable to communicate with the medical staff due to prior neurological deficits unrelated to the current event or was unable to communicate due to external circumstances such as mechanical ventilation. The exclusion criteria were: patients with a permanent urinary catheter, asymptomatic bacteriuria, and patients with other suspected sources of infection. During the study period, 543 single pathogen urine cultures were identified, 198 of which failed to meet the inclusion criteria (Figure 1). Of the remaining 345 patients, 122 cases presented with a positive culture obtained more than 48 hours following admission, therefore they could not be classified as community-acquired infection and were thus classified as possible hospital-acquired infections and excluded from the study. According to current local recommendations, a quinolone (IV ofloxacin, 200 mg bid) or an extended spectrum cephalosporin (IV cefuroxime, 750 mg tid), were used in our institution as the empiric first-line antibiotics for hospitalized CAUTI (presumed to be the more severe cases of CAUTI/ pyelonephritis). 9,11 Study patients were identified from positive urinary culture tests; study-criteria were next confirmed by comprehensive chart review as well as by patient interview during the period of hospitalization. Data collected included demographic details, underlying diseases, symptoms and signs related to the present infection, diagnosis upon admission, and initial antibiotic therapy administered within the first 24 hours of admission. Etiological pathogens and their sensitivity to antibiotic therapy were then evaluated. Urine cultures were carried out by our Microbiology Laboratory Service according to standard techniques. 15 Antimicrobial susceptibility tests were performed using the Kirby Bauer disk diffusion method 16 and interpretation of these tests was according to the guidelines proposed by the National Committee for Clinical Laboratory Standards. 17 The assay included eight antibiotics: ampicillin (Amp), ofloxacin (Oflo), amoxicillin/clavulanate (AM/CL), cefuroxime (Cef), gentamicin (Gen), nalidixic acid (Nal), trimethoprim/sulfamethoxazole (TMP/SMX), and nitrofurantoin (Nit). Enteric Gramnegative flora were defined as pathogens from one of the following identified species: Escherichia coli, Proteus mirabilis, Morganella morganii, Pseudomonas aeruginosa, Klebsiella spp, and Enterobacter spp. This was an observational study and the authors did not interfere with clinical decisions or with the selection of antibiotics. The study protocol was approved by the Institutional Helsinki Ethical Committee. Data analysis Bivariate hypotheses involving continuous variables were tested using the Student s t test for independent groups with normal distribution. Normality of the study data was tested Table 1 Diagnosis of UTI on admission to the Soroka Medical Center during 2000 All cases n = 223 Diagnosis on admission p Value UTI n = 54 Other n = 169 Female gender, (%) 146 (65.5) 34 (63.0) 112 (66.3) 0.66 Age, years 64.9 19.2 62.9 22.3 65.6 18.0 0.37 Recurrent, (%) 64 (28.7) 31 (57.4) 33 (19.5) <0.001 Previous antibiotic therapy, (%) 25 (11.2) 8 (14.8) 17 (10.1) 0.33 Dementia, (%) 47 (21.1) 12 (22.2) 35 (20.7) 0.81 Nursing home, (%) 22 (9.9) 10 (18.5) 12 (7.1) 0.01 Symptoms a Burning sensation, (%) 81 (46.0) 31 (73.8) 50 (37.3) <0.001 Urgency to urinate, (%) 73 (41.5) 25 (59.5) 48 (35.8) 0.01 Urinate frequently, (%) 51 (29.0) 15 (35.7) 36 (26.9) 0.27 Flank pain, (%) 73 (41.5) 22 (52.4) 51 (38.1) 0.09 Complicated UTI, (%) 126 (56.5) 32 (59.3) 94 (55.6) 0.64 Criteria for complication Urinary tract problems, (%) 70 (31.4) 19 (35.2) 51 (30.2) 0.49 Diabetes, (%) 57 (25.6) 14 (25.9) 43 (25.4) 0.94 Other, (%) 9 (4.0) 1 (1.9) 8 (4.7) 0.69 Antibiotic therapy First-line, (%) b 95 (42.6) 42 (77.8) 53 (31.4) <0.001 Other antibiotics, (%) 49 (22.0) 8 (14.8) 41 (24.3) No antibiotics, (%) 79 (35.4) 4 (7.4) 75 (44.4) a Symptoms were assessed only in 176 patients who were able to communicate. b Cefuroxime or ofloxacin were considered first-line antibiotics.

248 L. Nesher et al. Table 2 Baseline characteristics of the patient population according to the decision for antibiotic therapy upon admission Any antibiotic treatment n = 144 No antibiotics n = 79 p Value Female gender, (%) 89 (61.8) 57 (72.2) 0.12 Age, years 63.9 20.1 66.8 17.3 0.28 Recurrent, (%) 51 (35.4) 13 (16.5) 0.003 Previous antibiotic therapy, (%) 22 (15.3) 3 (3.8) 0.01 Dementia, (%) 27 (18.8) 20 (25.3) 0.25 Nursing home, (%) 14 (9.7) 8 (10.1) 0.92 Complicated UTI, (%) 84 (58.3) 42 (53.2) 0.46 Criteria for complication Urinary tract problems, (%) 56 (38.9) 14 (17.7) 0.001 Diabetes, (%) 27 (18.8) 30 (38.0) 0.002 Other, (%) 7 (4.9) 2 (2.5) 0.50 with a 1-sample Kolmogorov Smirnov test to indicate the appropriateness of parametric testing. For tests of whether the distribution of categorical variables differed across study groups the Chi-square test was used. Logistic regression models were used for multivariate analysis. All the variables found to correlate significantly with outcome in univariate analysis with p value less than 0.1 were included in the multivariate analysis. In order to minimize potential confounders, all variables with uneven distribution between two groups were also included in the multivariate analysis. Results The study included 223 patients; only 54 (24.2%) were diagnosed as having a CAUTI upon admission while 155 (69.5%) were so diagnosed on discharge. The demographic data and clinical information according to the diagnosis upon admission are presented in Table 1. As compared with patients diagnosed as UTI subsequent to admission, the group of patients diagnosed as CAUTI upon admission had a significantly higher rate of recurrence and complained more frequently of a burning sensation and a need to urinate. Of the 223 patients only 176 (78.9%) were able to communicate; failure to discuss complaints and symptoms was usually associated with dementia within the group of older CAUTI patients. Within this group, 42 patients diagnosed as CAUTI on admission had a median of two complaints as compared with one complaint in patients not diagnosed on admission ( p < 0.001). Of the 126 patients fulfilling the criteria for complicated UTI, only 25.4% were so diagnosed on admission. Although the fraction of patients from nursing homes was only 9.9% of the study population, within this subpopulation, CAUTI was more readily recognized in these patients ( p < 0.01). Patients diagnosed upon admission with CAUTI received first-line antibiotic therapy at a higher rate than non-first-line therapy (77.8% vs. 14.8%, p < 0.001). Multivariate analysis (logistic regression) revealed four factors associated with early diagnosing of CAUTI (upon admission): (1) recurrence (odds ratio 3.4; 95% CI 1.5 7.6); (2) burning sensation (odds ratio 4.4; 95% CI 1.9 10.2); (3) urgency to urinate (odds ratio 2.8; 95% CI 1.2 6.4); (4) flank pain (odds ratio 4.8; 95% CI 2.0 11.7). Data in Table 2 outline the patient population stratified by the decision made by the admitting physician to start administering antibiotic treatment. Of the 223 cases studied, 95 (42.6%) received a first-line antibiotic therapy for CAUTI recommended in our institution (cefuroxime or ofloxacin); an additional 49 (22.0%) received a different type of antibiotic and 79 (35.4%) did not receive any antibiotic therapy in the first 24 hours of hospitalization. As might be expected, a history of UTI or CAUTI could be seen as a major contributing factor affecting the decision of the admitting physician to initiate antibiotic drug therapy. Figure 2 shows the breakdown of the pathogens found in urinary cultures of the patients included in this study. As expected, Gram-negative intestinal flora comprised 86.1% (192 patients) of the offending microorganisms. Patients with Gram-negative enteric pathogens were somewhat older (66.2 19.1 vs. 56.9 18.0 years, p = 0.01) and had a higher rate of recurrence (30.7% vs. 16.1%, p = 0.09). Pathogen sensitivities in this study are shown in Table 3. The gastrointestinal tract is the most common source of pathogens in UTI; of the 192 cases of enteric pathogen cultures, 20.3% demonstrated resistance to ofloxacin, 19.8% were resistant to cefuroxime, and 15.6% demonstrated resistance to gentamicin. Enteric pathogen cultures resistant to both ofloxacin and cefuroxime were found in 19 cases (9.9%). As compared with the rest of the cohort, the subgroup Figure 2 Classification of pathogens on urine culture. Relative distribution of microorganisms identified in 223 cases of hospital UTI at the Soroka Medical Center during 2000. Other microorganisms included Acinetobacter (6), Enterobacter (9), coagulase negative staphylococci (3), Morganella (4), Strep B (1), and Citrobacter (1).

Regional community-acquired UTI in Israel 249 Table 3 Antibiotic sensitivity of the bacteria identified in the urine cultures Sensitivity Overall n = 223 E. coli n = 125 Pseudomonas n =15 Enterococcus n =20 Klebsiella n =23 Proteus n =16 Other a n =24 Cefuroxime 157 (70.4) 114 (91.2) Not tested Not tested 18 (78.3) 12 (75.0) 12 (50.0) Ofloxacin 157 (70.4) 102 (81.6) 10 (66.7) Not tested 17 (73.9) 12 (75.0) 16 (66.7) Gentamicin 177 (79.4) 108 (86.4) 1 (6.7) Not tested 17 (73.9) 7 (43.8) 18 (75.0) TMP/SMX 115 (51.6) 72 (57.6) 1 (6.7) Not tested 17 (73.9) 7 (43.8) 18 (75.0) TMP/SMX, trimethoprim/sulfamethoxazole. a Other microorganisms included Acinetobacter (6), Enterobacter (9), coagulase negative staphylococci (3), Morganella (4), Strep B (1), and Citrobacter (1). Table 4 Characteristics of the patient population according to sensitivity of bacteria to first-line antibiotics Sensitive to first-line antibiotics n = 180 Resistant to first-line antibiotics n =43 Female gender, (%) 123 (68.3) 23 (53.5) 0.07 Age, years 64.9 19.0 65.1 20.2 0.96 Recurrent, (%) 52 (28.9) 12 (27.9) 0.89 Previous antibiotic therapy, (%) 18 (10) 7 (16.3) 0.24 Dementia, (%) 36 (20) 11 (25.6) 0.42 Nursing home, (%) 15 (8.3) 7 (16.3) 0.11 Complicated UTI, (%) 97 (53.9) 29 (67.4) 0.10 Complications Urinary pathway problems, (%) 55 (30.6) 15 (34.9) 0.58 Diabetes, (%) 47 (26.1) 10 (23.3) 0.70 Other, (%) 6 (3.3) 3 (7.0) 0.38 p Value of patients with double antibiotic resistance was older (75.7 16.0 vs. 65.2 19.1 years, p = 0.02), predominantly male (57.9% vs. 30.1%, p = 0.02), and most were nursing home residents (36.8% vs. 8.7%, p < 0.001). Within the subgroup of 20 cases of CAUTI caused by Enterococcus, all microorganisms were sensitive to vancomycin, 18 (90%) were ampicillin sensitive, and nine (45%) were sensitive to gentamicin. As can be seen in Table 4, no significant difference in resistance to the two first-line antibiotics was observed between CAUTI and hospital-acquired UTI: in 29 out of 126 patients (23%) with complicated CAUTI, we identified pathogens with resistance to both first-line antibiotics, whereas 14 of 97 cultures (14.4%) of the non-complicated CAUTI showed resistance to both first-line antibiotics ( p = 0.10). Discussion When assessing the data collected in the present study, three major issues emerged: 1. We observed a low rate of diagnosis of CAUTI upon admission and prior to receiving a positive urine culture result. 2. When examining the group as a whole, the overall use of a first-line antibiotic upon admission was low; however, in cases diagnosed upon admission as CAUTI, the use of a first-line antibiotic was significantly higher. 3. A substantial percentage of the pathogens identified in our study population were resistant to first-line antibiotics, significantly more so in patients with complicated CAUTI. Only 24.2% of the CAUTI were identified prior to a positive urine culture result, despite the fact that most of these patients presented with fever and urinary tract symptoms. More importantly, only 69.5% of cases were diagnosed and listed as CAUTI upon discharge. The latter can partly be attributed to early discharge, prior to the return of the urinary culture result. One may predict that the introduction of automation into analysis and on-line data transfer would most likely improve the rate of early identification of CAUTI. Patients diagnosed as CAUTI upon admission received a first-line antibiotic at a much higher rate (77.8%), than the group of undiagnosed patients (31.4%). This clearly suggests that inadequate diagnosis upon admission is to blame rather than staff failure to comply with the guidelines. Considering the entire case record covered in this study, the first-line CAUTI antibiotics (ofloxacin or cefuroxime) appear inappropriate for 14% of the non-complicated CAUTI. We observed an even higher rate of resistance to either one of the first-line antibiotics in the group of complicated CAUTI: cultures of 23% of the patients with complicated CAUTI were resistant to both first-line antibiotics. Therefore, this study suggests that, as far as our center and our region is concerned, patients presenting with complicated CAUTI failed to receive adequate antibiotic coverage in almost 25% of admissions. A major incentive for this study was to monitor regional changes in UTI pathogens. Thus, and despite the obvious reservation for protocol differences, a study done in this region in 1995, 18 already pointed to a trend of decline in Gram-negative enteric pathogens in urine cultures of suspected UTI patients in our community clinics. Over a period

250 L. Nesher et al. of four years, that study reported a significant decline in E. coli positive cultures (70.5% to 61.2%) and a decline in Proteus mirabilis, Morganella morganii, and Pseudomonas aeruginosa and a rise in frequency of Klebsiella spp and Enterobacter spp. Also, resistance to first line antibiotics has increased in this study as compared with the report by Weber et al. 18 Thus, probable regional changes in pathogens and in their sensitivities to first line antibiotics should clearly be the motivation for periodical analyses such as this. Patients who were included in this study were mostly from the severe end of the spectrum of CAUTI including pyelonephritis and urosepsis, since patients suffering from simple cystitis or uncomplicated CAUTI or young patients with acute pyelonephritis are usually managed on an outpatient basis. One clear limitation of this study relates to its observational nature; this may have caused some cases to be missed or misclassified as CAUTI. As many as 47 patients suffered from cognitive impairment that hindered the communication of their symptoms. This group of patients was included in the study after alternative sources of infection were ruled out. However, these patients were excluded from the specific analysis of symptoms and complaints. In summary: 1. Resistance to first-line antibiotics was more common among males than females (38% vs. 8%). 2. A higher rate of resistance to antibiotics was observed in complicated UTI as compared to non-complicated UTI (24% vs. 15%, relative risk 1.6). 3. In our institution, only a small percentage of the UTI patients were treated with a first-line antibiotic; we believe that this was due to failure of appropriate diagnosis of CAUTI upon admission. We observed a trend of increased antibiotic resistance in our local pathogens, specifically in males and in complicated CAUTI patients. To meet such changes, similar studies should be repeated at regular intervals. 4. A combination of two antibiotics should be considered in cases of complicated UTI and in male patients with CAUTI. 5. Prompt diagnosis assists in providing adequate antibiotic therapy, which may reduce in-hospital mortality from CAUTI. Conflict of interest: No conflict of interest to declare. References 1. Foxman B. Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Am J Med 2002;113(Suppl 1A): 5S 13S. 2. Barnett BJ, Stephens DS. Urinary tract infection: an overview. Am J Med Sci 1997;314:245 9. 3. Orenstein R, Wong ES. Urinary tract infections in adults. Am Fam Physician 1999;59:1225 34. 37. 4. Hooton TM, Scholes D, Hughes JP, Winter C, Roberts PL, Stapleton AE, et al. A prospective study of risk factors for symptomatic urinary tract infection in young women. N Engl J Med 1996; 335:468 74. 5. Gupta K, Hooton TM, Stamm WE. Increasing antimicrobial resistance and the management of uncomplicated communityacquired urinary tract infections. Ann Intern Med 2001;135: 41 50. 6. Manges AR, Johnson JR, Foxman B, O Bryan TT, Fullerton KE, Riley LW. Widespread distribution of urinary tract infections caused by a multidrug-resistant Escherichia coli clonal group. N Engl J Med 2001;345:1007 13. 7. Colodner R, Keness Y, Chazan B, Raz R. Antimicrobial susceptibility of community-acquired uropathogens in northern Israel. Int J Antimicrob Agents 2001;18:189 92. 8. Huang ES, Stafford RS. National patterns in the treatment of urinary tract infections in women by ambulatory care physicians. Arch Intern Med 2002;162:41 7. 9. Warren JW, Abrutyn E, Hebel JR, Johnson JR, Schaeffer AJ, Stamm WE. Guidelines for antimicrobial treatment of uncomplicated acute bacterial cystitis and acute pyelonephritis in women. Infectious Diseases Society of America (IDSA). Clin Infect Dis 1999;29:745 58. 10. Elhanan G, Sarhat M, Raz R. Empiric antibiotic treatment and the misuse of culture results and antibiotic sensitivities in patients with community-acquired bacteraemia due to urinary tract infection. J Infect 1997;35:283 8. 11. Rubenstein JN, Schaeffer AJ. Managing complicated urinary tract infections: the urologic view. Infect Dis Clin North Am 2003;17:333 51. 12. Pinson AG, Philbrick JT, Lindbeck GH, Schorling JB. Fever in the clinical diagnosis of acute pyelonephritis. Am J Emerg Med 1997;15:148 51. 13. Hooton TM, Stamm WE. Diagnosis and treatment of uncomplicated urinary tract infection. Infect Dis Clin North Am 1997; 11:551 81. 14. Saint S, Scholes D, Fihn SD, Farrell RG, Stamm WE. The effectiveness of a clinical practice guideline for the management of presumed uncomplicated urinary tract infection in women. Am J Med 1999;106:636 41. 15. Pezzlo M. Clinical microbiology procedures handbook. Washington, D.C., USA: American Society for Microbiology; 1992. 16. Bauer AW, Kirby WM, Sherris JC, Turck M. Antibiotic susceptibility testing by a standardized single disk method. Am J Clin Pathol 1966;45:493 6. 17. NCCLS. Performance standards for antimicrobial disk susceptibility tests. Fifth edition; approved standard. NCCLS document M2-A5. Villa-Nova, Pennsylvania, USA: NCCLS; 1993. 18. Weber G, Riesenberg K, Schlaeffer F, Peled N, Borer A, Yagupsky P. Changing trends in frequency and antimicrobial resistance of urinary pathogens in outpatient clinics and a hospital in Southern Israel, 1991 1995. Eur J Clin Microbiol Infect Dis 1997;16: 834 8.